A seawater desalination load forecasting method and system

Abstract

The present invention relates to a method and system for predicting seawater desalination load, comprising: based on pre-set influencing factors, obtaining historical feature vectors and loads with the greatest similarity to the prediction day from historical data, bringing the feature vectors and loads into a pre-established seawater desalination load prediction model, predicting the load value on the prediction day, and obtaining the load prediction value on the prediction day. The present invention relates to a method and system for predicting seawater desalination load, comprising: based on pre-set influencing factors, obtaining historical feature vectors and loads with the greatest similarity to the prediction day from historical data, bringing the feature vectors and loads into a pre-established seawater desalination load prediction model, predicting the load value on the prediction day, and obtaining the load prediction value on the prediction day, reducing the power consumption and overall cost of seawater desalination, and also accurately predicting the seawater desalination load to achieve the purpose of coordinated operation with distributed renewable energy or power grid.

Description

Technical Field

[0001] The present invention relates to the field of seawater desalination load regulation, and in particular to a seawater desalination load prediction method and system. Background Art

[0002] As a solution to the shortage of fresh water resources in my country, seawater desalination has become an important load in coastal areas. Traditional seawater desalination consumes a lot of electricity, has a high overall cost, and has a slow progress in industrial promotion. In order to further improve the comprehensive economic benefits of seawater desalination, the seawater desalination load can be combined with distributed renewable energy or power grids to participate in regulation, and the cost can be reduced by reducing the electricity consumption of the seawater desalination load. Summary of the invention

[0003] In order to solve the above-mentioned problems of large power consumption, high overall cost and slow industrial promotion progress of traditional seawater desalination, the present invention provides a seawater desalination load forecasting method, the method comprising:

[0004] Based on the pre-set influencing factors, the historical feature vector and load with the greatest similarity to the forecast day are obtained from the historical data;

[0005] Bringing the characteristic vector and the load into a pre-established seawater desalination load forecasting model, forecasting the load value on the forecast day, and obtaining the load forecast value on the forecast day;

[0006] The seawater desalination load prediction model is constructed based on a least squares support vector machine with particle swarm optimization regularization parameters and kernel parameters.

[0007] Preferably, the influencing factors include: temperature, daily water consumption, seawater flow rate, fresh water return flow rate, high-pressure brine flow rate, feed pump, high-pressure pump and booster pump timing power, return pump timing power.

[0008] Preferably, the step of obtaining the historical feature vector with the greatest similarity to the predicted day from the historical data further includes:

[0009] Normalizing the historical data respectively;

[0010] Construct factor sequences based on normalized historical data;

[0011] Based on the factor sequence, a feature vector with the greatest similarity to the predicted day is determined.

[0012] Preferably, the step of constructing a factor sequence based on normalized historical data includes:

[0013] Obtaining a first characteristic value based on the normalized temperature data;

[0014] Obtaining a second characteristic value based on the normalized daily water consumption data;

[0015] The third characteristic quantity is obtained based on the time-series power consumption of the water feed pump, the high-pressure pump and the booster pump of the seawater desalination unit after normalization processing;

[0016] The fourth characteristic quantity is obtained based on the time-series power consumption of the return pump of the seawater desalination unit after normalization;

[0017] A fifth characteristic value is obtained based on the normalized high-pressure concentrated brine flow rate of the seawater desalination unit;

[0018] A factor sequence is constructed by the first feature amount, the second feature amount, the third feature amount, the fourth feature amount, and the fifth feature amount.

[0019] Preferably, determining the feature vector with the greatest similarity to the predicted day based on the factor sequence includes:

[0020] Based on the factor sequence, determining the difference sequence between the historical data and the forecast day data;

[0021] Obtaining a maximum difference and a minimum difference based on the difference sequence;

[0022] Based on the maximum difference and the minimum difference, the grey correlation degree is used to calculate the grey correlation degree of the predicted day factor to the historical day factor;

[0023] Based on the grey correlation degree, multiple historical day factors with the greatest similarity to the predicted day are obtained;

[0024] Take all historical day factors as feature vectors.

[0025] Preferably, the step of bringing the characteristic vector and the load into a pre-established seawater desalination load forecasting model to forecast the load value on the forecast day comprises:

[0026] The seawater desalination load forecasting model is initialized by least squares support vector machine based on particle swarm optimization regularization parameters and kernel parameters;

[0027] dividing the feature vectors and the loads into training data and test data;

[0028] Based on the training data, the total number of feature vectors is input into the seawater desalination load prediction model, the seawater desalination load prediction model is trained, and the trained seawater desalination load prediction value is obtained;

[0029] Based on the test data, the characteristic vector is input into the seawater desalination load forecasting model to obtain the load forecasting data;

[0030] Based on the load corresponding to the test data, the load prediction data is corrected to obtain a seawater desalination load prediction value.

[0031] Preferably, the step of inputting the characteristic vector into the seawater desalination load forecasting model based on the test data to obtain the load forecasting data comprises:

[0032] Based on the test data, the trained seawater desalination model is tested using a relative error method;

[0033] Bringing the test data into the seawater desalination model to obtain a seawater desalination load prediction value;

[0034] Constructing a relative error function based on the predicted value of seawater desalination and the actual measured value of seawater desalination corresponding to the test data;

[0035] The relative error function is used as the objective function, and the objective function is optimized using a particle swarm algorithm to obtain an optimized seawater desalination load prediction value.

[0036] Preferably, the relative error function is as follows:

[0037]

[0038] In the formula, y real Indicates the actual value of seawater load; y predicted Indicates the predicted value of seawater load.

[0039] Preferably, the predicted load model is as follows:

[0040]

[0041] Where y represents the predicted value of the desalination load; i represents the i-th characteristic quantity; t represents the total number of characteristic quantities; λ i represents the Lagrange multiplier of the i-th characteristic quantity; x i represents the vector composed of the i-th feature quantity; x represents the vector composed of the feature quantity; b represents the offset; K(x i ,x) represents the kernel function;

[0042] The formula of the kernel function is as follows:

[0043] K(x i ,x)=exp{-||x i -x|| 2 / 2σ 2}

[0044] In the formula, σ 2 Represents the radial basis function kernel width.

[0045] A seawater desalination load forecasting system, comprising:

[0046] Acquisition module: used to obtain the historical feature vector and load with the greatest similarity to the forecast day from historical data based on pre-set influencing factors;

[0047] Prediction module: used for bringing the characteristic vector and load into a pre-established seawater desalination load prediction model, predicting the load value on the prediction day, and obtaining the load prediction value on the prediction day;

[0048] The seawater desalination load prediction model is constructed based on a least squares support vector machine with particle swarm optimization regularization parameters and kernel parameters.

[0049] Preferably, the prediction module includes a prediction unit;

[0050] The prediction unit is used to determine an initialized seawater desalination load prediction model based on a least squares support vector machine of particle swarm optimization regularization parameters and kernel parameters;

[0051] dividing the feature vectors and the loads into training data and test data;

[0052] Based on the training data, the total number of feature vectors is input into the seawater desalination load prediction model, the seawater desalination load prediction model is trained, and the trained seawater desalination load prediction value is obtained;

[0053] Based on the test data, the characteristic vector is input into the seawater desalination load forecasting model to obtain the load forecasting data;

[0054] Based on the load corresponding to the test data, the load prediction data is corrected to obtain a seawater desalination load prediction value.

[0055] Compared with the closest prior art, the present invention has the following beneficial effects:

[0056] 1. The present invention is a method and system for predicting seawater desalination load. Based on pre-set influencing factors, the historical feature vector and load with the greatest similarity to the prediction day are obtained from historical data, and the feature vector and load are brought into a pre-established seawater desalination load prediction model to predict the load value of the prediction day, thereby obtaining the load prediction value of the prediction day, thereby reducing the seawater desalination power consumption and overall cost, and accurately predicting the seawater desalination load to achieve the purpose of coordinated operation with distributed renewable energy or power grid;

[0057] 2. The present invention is a method and system for predicting seawater desalination load. The proposed prediction model has good stability, generalization ability and reliable prediction performance, which can meet the needs of seawater desalination load prediction. At the same time, the prediction results can also provide reference and decision support for seawater desalination to participate in power grid or distributed renewable energy peak regulation and even frequency regulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0059] Figure 2 This is a diagram of a seawater desalination device according to the present invention;

[0060] Figure 3 This is a flow chart of the seawater desalination load prediction of the present invention. Specific embodiments

[0061] The present invention will be further explained below in conjunction with specific embodiments:

[0062] Example 1

[0063] like Figure 1 As shown: A method for predicting seawater desalination load, the method comprising:

[0064] Step 1: Based on the pre-set influencing factors, obtain the historical feature vector and load with the greatest similarity to the forecast day from the historical data;

[0065] Step 2: Bring the characteristic vector and the load into a pre-established seawater desalination load forecasting model, forecast the load value on the forecast day, and obtain the load forecast value on the forecast day;

[0066] The seawater desalination load prediction model is constructed based on a least squares support vector machine of particle swarm optimization related parameters.

[0067] In order to further illustrate the technical means adopted by the present invention and the effects achieved, the technical solution of the present invention is clearly and completely described below in conjunction with the accompanying drawings and specific embodiments;

[0068] The explanation of step 1: based on the pre-set influencing factors, obtaining the historical feature vector and load with the greatest similarity to the forecast day from the historical data is as follows:

[0069] like Figure 2 , is a flow chart of a method for predicting seawater desalination load based on particle swarm optimization least squares support vector machine of the present invention. The prediction method of the present invention comprises the following steps:

[0070] S101 obtains historical data related to seawater desalination load forecasting, including: temperature data, daily water consumption data, seawater flow rate data, freshwater return flow rate data, high-pressure concentrated brine flow rate data, feed pump, high-pressure pump and booster pump time-series power data, return pump time-series power data;

[0071] S102 analyzes and processes the historical data, and the processing process is as follows:

[0072] 1) Normalize the temperature data as feature 1;

[0073] 2) Taking the mapping value of daily water consumption data as feature 2;

[0074] 3) Mapping the time series of electric power used by the desalination unit feed water pump, high pressure pump and booster pump to the feed water pump seawater flow rate data sequence, and normalizing them as feature quantity 3;

[0075] 4) Mapping the time series of electric power used by the desalination unit return pump to the fresh water return flow rate data series, and performing normalization processing as feature quantity 4;

[0076] 5) Mapping the high-pressure concentrated brine flow rate data of the seawater desalination unit to the fresh water return flow rate data, and normalizing them as feature quantity 5;

[0077] 6) Feature quantity 1, feature quantity 2, feature quantity 3, feature quantity 4 and feature quantity 5 are used as feature quantities for similar day selection.

[0078] S103 selects similar days and constructs a small sample set of similar days based on relevant factors using grey correlation theory.

[0079] The difference between the historical day factors and the day factors to be predicted:

[0080] The factor sequence X of historical days n m =(x m (1),…,x m (t))

[0081] The factor sequence of the day to be predicted is X0 = (x0(1),…,x0(t))

[0082] Where m = 1 , 2,…n; t is the number of factors taken into account;

[0083] The difference of the sequence Δ=[Δ 10 ,…,Δ m0 ,…,Δ n0 ], where Δ m0 =(|x m (1)-x0(1)|,…,|x m (t)-x0(t)|)

[0084] The maximum and minimum difference between the two levels:

[0085]

[0086]

[0087] X0 vs X m Grey relational degree:

[0088]

[0089] Among them, Δ m0 (k)=|x m (k)-x0(k)|; ρ is the resolution coefficient, which is 0.5.

[0090] For step 2: bringing the characteristic vector and the load into the pre-established seawater desalination load forecasting model, forecasting the load value on the forecast day, and obtaining the load forecast value on the forecast day, the explanation is as follows:

[0091] Using the similar day data, a seawater desalination load prediction model based on a least squares support vector machine with particle swarm optimization related parameters is established;

[0092] As an embodiment, the seawater desalination load prediction model based on particle swarm optimization least squares support vector machine is a three-layer structure;

[0093] The first layer of the structure of the seawater desalination load forecasting model based on particle swarm optimization least squares support vector machine is the input layer. The input variables are the mapping values ​​of several daily water consumption, seawater flow rate, freshwater return flow rate and the temperature of similar days, high-pressure concentrated brine flow rate, feed water pump, high-pressure pump and booster pump sequential power, and return water pump sequential power vectors related to load forecasting. The input variables form a vector x i , where i = 1, 2…l;

[0094] l is the number of input variables in the input layer;

[0095] The second layer of the model structure is a training layer, which maps high-dimensional input samples from the original space to a high-dimensional feature space through nonlinear transformation, and uses the training samples in this space to train the seawater desalination load prediction model of the least squares support vector machine;

[0096] The optimal linear regression function constructed in this space:

[0097]

[0098] According to statistical theory, LS-SVM determines the regression function by minimizing the following target numbers:

[0099]

[0100]

[0101] In the above formula, e i is the error; γ is the regularization parameter, which controls the degree of penalty for the error. By introducing the Lagrange multiplier, the above formula is transformed into:

[0102]

[0103] Taking all partial derivatives to zero, we get

[0104]

[0105] After eliminating ω and e, the optimal problem is transformed into a linear system to be solved. λ can be obtained by solving the linear equation i and b, so the expression of the linear prediction model is:

[0106]

[0107] in, Refers to the kernel function. The radial basis function (RBF) is the most commonly used expression for calculating the kernel function. The equation of the kernel function is:

[0108] K(x i ,x)=exp{-||x i -x|| 2 / 2σ 2}

[0109] where σ 2 is the RBF kernel tuning parameter and represents the RBF kernel width.

[0110] The third layer of the model is the output layer. The output layer uses the regression function after training based on the least squares support vector machine prediction model as the output basis. The output is recorded as y i , where i=1,2,…l is the corresponding number of samples;

[0111] S105 uses a relative error (RE) method based on the load forecast data of several forecast days and the actual data of the forecast day to evaluate the model prediction ability in the training and testing stages. The calculation of the RE is as follows:

[0112]

[0113] S106 uses the relative error function as the objective function, and uses the particle swarm algorithm to optimize the parameters of the objective function. After the optimization is completed, a seawater desalination load prediction model based on the least squares support vector machine optimized by the particle swarm is established;

[0114] As an embodiment, before optimizing the objective function using the particle swarm optimization algorithm, the parameters of the seawater desalination load prediction model based on the least squares support vector machine that need to be determined must be indicated;

[0115] The seawater desalination load prediction model based on least squares support vector machine needs to select two model parameters, namely, regularization parameter γ and kernel parameter σ 2;

[0116] like Figure 3 As shown in Figure 1, the particle swarm optimization algorithm is a very simple algorithm that can effectively optimize various functions. PSO has the advantages of simple concept, easy implementation and fast convergence. The optimization of the objective function parameters based on the particle swarm optimization algorithm includes the following steps:

[0117] 1) Initialize the parameters of the least squares support vector machine model, set the regularization parameter γ and the kernel parameter σ 2 The value range of

[0118] 2) Assign and initialize individual extreme value positions P i (t) = [p i,1 (t),p i,2 (t),…,p i,d (t)] and its corresponding individual extreme value is called the local optimum (pbest), the global extreme position P g (t) = [p g,1 (t),p g,2 (t),…,p g,d (t)] and its corresponding global extreme value are called global optimum (gbest);

[0119] 3) Calculate the fitness p of each particle f ;

[0120] First, output a particle. For each sample, the output value of a model can be calculated, and then the relative error is calculated according to the fitness function. In the same way, the relative error of all samples is calculated, which is the fitness of the corresponding particle. Continue to input other particles until the fitness of all particles is calculated.

[0121] 4) The particle updates its position and velocity by tracking two "extreme values" (pbest, gbest). The new velocity and position of the particle are updated by the following equations:

[0122] v i，j (t+1)=ωv i，j (t)+c1r1(p i，j -x i，j (t))+c2r2(p g，j -x i，j (t))

[0123] x i,j (t+1)=x i,j (t)+v i,j (t+1)

[0124] Where ω is the inertia weight factor; r1 and r2 are two independent random numbers in the range [0,1]; c1 and c2 are two positive constants, called acceleration factors.

[0125] 5) Compare whether the number of iterations reaches the maximum number of iterations or the preset accuracy. If the preset accuracy is met or the maximum number of iterations is reached, the algorithm converges and the calculation ends; otherwise, return to step 7.3 and continue iterating until the maximum number of iterations or the preset accuracy is reached and the calculation ends.

[0126] S107 predicts the load of the prediction day according to the least squares support vector machine seawater desalination load prediction model based on particle swarm optimization;

[0127] Inputting historical data of several historical days into the established least squares support vector machine seawater desalination load forecasting model based on particle swarm optimization, the output is the load forecast value of the day to be forecasted;

[0128] Compared with general technologies, the present invention considers factors that have a greater impact on the load, such as daily water consumption, seawater flow rate, freshwater return flow rate and temperature, and high-pressure concentrated brine flow rate, in the seawater desalination load forecasting. The algorithm has a complete theoretical basis, and the model has good stability, generalization ability and reliable prediction performance, which can meet the needs of seawater desalination load forecasting. At the same time, the prediction results can also provide reference and decision support for seawater desalination to participate in the peak load regulation or even frequency regulation of the power grid or distributed renewable energy.

[0129] Example 2

[0130] The present invention also relates to a seawater desalination load forecasting system, comprising:

[0131] Acquisition module: used to obtain the historical feature vector and load with the greatest similarity to the forecast day from historical data based on pre-set influencing factors;

[0132] Prediction module: used for bringing the characteristic vector and load into a pre-established seawater desalination load prediction model, predicting the load value on the prediction day, and obtaining the load prediction value on the prediction day;

[0133] The seawater desalination load prediction model is constructed based on a least squares support vector machine with particle swarm optimization regularization parameters and kernel parameters.

[0134] The prediction module includes a prediction unit;

[0135] The prediction unit is used to determine an initialized seawater desalination load prediction model based on a least squares support vector machine of particle swarm optimization regularization parameters and kernel parameters;

[0136] dividing the feature vectors and the loads into training data and test data;

[0137] Based on the training data, the total number of feature vectors is input into the seawater desalination load prediction model, the seawater desalination load prediction model is trained, and the trained seawater desalination load prediction value is obtained;

[0138] Based on the test data, the characteristic vector is input into the seawater desalination load forecasting model to obtain the load forecasting data;

[0139] Based on the load corresponding to the test data, the load prediction data is corrected to obtain a seawater desalination load prediction value.

[0140] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0141] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0142] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0143] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1The steps for the functions specified in one or more boxes.

[0144] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A method for predicting seawater desalination load, characterized in that: The method comprises: Based on the pre-set influencing factors, the historical feature vector and load with the greatest similarity to the forecast day are obtained from the historical data; Bringing the characteristic vector and the load into a pre-established seawater desalination load forecasting model, forecasting the load value on the forecast day, and obtaining the load forecast value on the forecast day; The seawater desalination load prediction model is constructed based on a least squares support vector machine with particle swarm optimization regularization parameters and kernel parameters; The influencing factors include: temperature, daily water consumption, seawater flow rate, freshwater return flow rate, high-pressure concentrated brine flow rate, feed pump, high-pressure pump and booster pump timing power, return pump timing power; The step of obtaining the historical feature vector with the greatest similarity to the predicted day from the historical data further includes: Normalizing the historical data respectively; Construct factor sequences based on normalized historical data; Determine the feature vector with the greatest similarity to the forecast day based on the factor sequence; The step of constructing a factor sequence based on normalized historical data includes: Obtaining a first characteristic value based on the normalized temperature data; Obtaining a second characteristic value based on the normalized daily water consumption data; The third characteristic quantity is obtained based on the time-series power consumption of the water feed pump, the high-pressure pump and the booster pump of the seawater desalination unit after normalization processing; The fourth characteristic quantity is obtained based on the time-series power consumption of the return pump of the seawater desalination unit after normalization; A fifth characteristic value is obtained based on the normalized high-pressure concentrated brine flow rate of the seawater desalination unit; A factor sequence is constructed by the first feature amount, the second feature amount, the third feature amount, the fourth feature amount, and the fifth feature amount.

2. A method for predicting seawater desalination load according to claim 1, characterized in that: Determining the feature vector with the greatest similarity to the predicted day based on the factor sequence includes: Based on the factor sequence, determining the difference sequence between the historical data and the forecast day data; Obtaining a maximum difference and a minimum difference based on the difference sequence; Based on the maximum difference and the minimum difference, the grey correlation degree is used to calculate the grey correlation degree of the predicted day factor to the historical day factor; Based on the grey correlation degree, multiple historical day factors with the greatest similarity to the predicted day are obtained; Take all historical day factors as feature vectors.

3. A method for predicting seawater desalination load according to claim 2, characterized in that: The step of bringing the characteristic vector and the load into a pre-established seawater desalination load forecasting model to forecast the load value on the forecast day includes: The seawater desalination load forecasting model is initialized by least squares support vector machine based on particle swarm optimization regularization parameters and kernel parameters; dividing the feature vectors and the loads into training data and test data; Based on the training data, the total number of feature vectors is input into the seawater desalination load prediction model, the seawater desalination load prediction model is trained, and the trained seawater desalination load prediction value is obtained; Based on the test data, the characteristic vector is input into the seawater desalination load forecasting model to obtain the load forecasting data; Based on the load corresponding to the test data, the load prediction data is corrected to obtain a seawater desalination load prediction value.

4. A method for predicting seawater desalination load according to claim 3, characterized in that: Based on the test data, inputting the characteristic vector into the seawater desalination load forecasting model to obtain the load forecasting data includes: Based on the test data, the trained seawater desalination model is tested using a relative error method; Bringing the test data into the seawater desalination model to obtain a seawater desalination load prediction value; Constructing a relative error function based on the predicted value of seawater desalination and the actual measured value of seawater desalination corresponding to the test data; The relative error function is used as the objective function, and the objective function is optimized using a particle swarm algorithm to obtain an optimized seawater desalination load prediction value.

5. A method for predicting seawater desalination load according to claim 4, characterized in that: The relative error function is as follows: In the formula, y real Indicates the actual value of seawater load; y predicted Indicates the predicted value of seawater load.

6. A method for predicting seawater desalination load according to claim 5, characterized in that: The load forecasting model is as follows: Where y represents the predicted value of seawater desalination load; i represents the i-th characteristic quantity; t represents the total number of features; λ i represents the Lagrange multiplier of the i-th characteristic quantity; x i represents the vector composed of the i-th feature quantity; x represents the vector composed of the feature quantity; b represents the offset; K(x i ,x) represents the kernel function; The formula of the kernel function is as follows: K(x i ,x)=exp{-||x i -x|| 2 / 2σ 2 } In the formula, σ 2 Represents the radial basis function kernel width.

7. A seawater desalination load forecasting system, characterized in that: include: Acquisition module: used to obtain the historical feature vector and load with the greatest similarity to the forecast day from historical data based on pre-set influencing factors; Prediction module: used for bringing the characteristic vector and load into a pre-established seawater desalination load prediction model, predicting the load value on the prediction day, and obtaining the load prediction value on the prediction day; The seawater desalination load prediction model is constructed based on a least squares support vector machine with particle swarm optimization regularization parameters and kernel parameters; The influencing factors include: temperature, daily water consumption, seawater flow rate, freshwater return flow rate, high-pressure concentrated brine flow rate, feed pump, high-pressure pump and booster pump timing power, return pump timing power; The acquisition module further includes: Normalizing the historical data respectively; Construct factor sequences based on normalized historical data; Determine the feature vector with the greatest similarity to the forecast day based on the factor sequence; The step of constructing a factor sequence based on normalized historical data includes: Obtaining a first characteristic value based on the normalized temperature data; Obtaining a second characteristic value based on the normalized daily water consumption data; The third characteristic quantity is obtained based on the time-series power consumption of the water feed pump, the high-pressure pump and the booster pump of the seawater desalination unit after normalization processing; The fourth characteristic quantity is obtained based on the time-series power consumption of the return pump of the seawater desalination unit after normalization; A fifth characteristic value is obtained based on the normalized high-pressure concentrated brine flow rate of the seawater desalination unit; A factor sequence is constructed by the first feature amount, the second feature amount, the third feature amount, the fourth feature amount, and the fifth feature amount.

8. A seawater desalination load forecasting system as claimed in claim 7, characterized in that: The prediction module includes a prediction unit; The prediction unit is used to determine an initialized seawater desalination load prediction model based on a least squares support vector machine of particle swarm optimization regularization parameters and kernel parameters; dividing the feature vectors and the loads into training data and test data; Based on the training data, the total number of feature vectors is input into the seawater desalination load prediction model, the seawater desalination load prediction model is trained, and the trained seawater desalination load prediction value is obtained; Based on the test data, the characteristic vector is input into the seawater desalination load forecasting model to obtain the load forecasting data; Based on the load corresponding to the test data, the load prediction data is corrected to obtain a seawater desalination load prediction value.

Images